Psuedo Random CMOS circuit

[loss1234]

check this out

Id love to get people talking about this


http://www.media.mit.edu/physics/pedagogy/fab/electronics/assignment.html

i built one and have been working on some modifications.

[ericcoleridge]

Loss,

This reminds me of something I was hoping to get people talking about in the Ken Stone/CGS forum.

I had been looking at Buchla Source of Uncertainty schematics (there is a model 265, 266, and now 266e), parts of which use shift registers, feedback, EXOR gates, etc. to provide various 'random' cvs. One thing that I really like about the Buchla circuits however, that may be exclusive to buchla, is that they give some control over the randomness of their outputs. They provide some selection over various voltage ranges, and time ranges.
Theres some description in the first link, and more schems in the second (also Doepher and Buchla both currently offer versions of the SOU, with good descriptions on the respective sites):

http://www.simple-answer.com/DIY.html

and

http://www.musicsynthesizer.com/Buchla/source_of_uncertainty.htm


Anyways, there are also a few Ken Stone designs, that aren't necessarily described as such, but may be similar to these buchla circuits, and other shift register based 'random' voltage circuits:

Infinite Melody, Gated Comparator, and Digital Noise

You may be interested in taking a look at some of this stuff, if you haven't already. It's been peeking my interest lately.

I hope this reply isn't too off topic/forum. I apologize if so.

[loss1234]

you know i have also been looking into these things. ITS JUST THE BUCHLA SCHEMATICS TERRIFY ME!!

i wish they were more modern looking, easier to read.

and btw, the lunetta area just feels right for some of these topics. non-traditional synth circuits that use lots of logic?? or am i wrong.

thanks for the links

id love to flesh this out some more maybe tomorrow
you know we STILL Need to hang out..we live close by!

[ericcoleridge]

[loss1234]

hey can you maybe mention the specific ken stone circuits you are speaking of? he has so many. i have looked at his digital noise quite a bit and have built his vco and vca.

thanks

[ericcoleridge]

The ones I linked to above are the ones I was talking about

The CGS Infinite Melody is already kinda close to the "Stored Random Voltage" section of the Buchla 266. Both use a series of shift registers (4015), with a feedback scheme going through an EXOR gate (4030). They both take a clock input and a CV input. What is generated, in both circuits, is a bunch of random numbers, or random voltage levels on each of a dozen or so register locations. These voltages are summed and sent to outputs.

One output is a purely random voltage, the other, more interesting output is random, but where the 'distribution' or range of voltage steps is, quoting Ken Stone, "inversely related to the frequency at which it will occur". It's 1 over F. The greater the pitch step, the less likely it is to occur.

In the CGS circuit, at least, this is done with a 4024 frequency divider, dividing down the clock rate of the shift registers. The shift register receiving the slowest frequency division, from a 4024 output, outputs the largest pitch steps, but less frequently. Each individual register output is limited by resistors, so, I think, the shift register chip clocked at the slowest rate has the smallest resistors, and therefore is allowed to output the largest voltages.

Anyways, in the Buchla 'Stored Random Voltage' section, there is also VC over this 'distribution'. On the CGS, I think there's just the two outputs without control over distribution.

Now, the other section of the Buchla 266, called "Quantized Random Voltage" is also done with a shift register (again, 4015s), and EXOR feedback. Again, the random voltages coming from each register location, are limited by certain resistor values, and then summed. One output, using resistors of equal value, gives a random, but equal distribution of "quantized" notes (sent to a VCO). The other output, the resistors again limit the output voltages to specific "quantized" levels, creating 64 possible voltage steps. You can then control the range of randomness by limiting the number of possible steps. This is done with another EXOR gate, and an input voltage.

The CGS module that could possibly be adapted to the Quantized Random Voltage circuit is the CGS Gated Comparator. On the CGS circuit, the shift register positions can be summed, but call for different resistor values before summing. Then it would be a matter of gating the output again. Or maybe not. There's probably more to it, but it seems like these circuits are pretty close to the Buchla circuits.


[/b]

[Scott Stites]

Actually the 266 SOU Quantized Random Voltage uses the CD4006, though CD4015s could be adapted for use by it.

Jeff Pontius and I went on a 266 SOU kick back in 2004. I breadboarded all of the circuits from it - what a power packed module. My two favorite functions from it were the Fluctuating Random Voltage and that funky, funky sample and hold. All of the functions were pretty cool though. Jeff actually built a module based off of it. I got sidetracked on the Quantized Random Voltage (the project turned into the first Klee sequencer, which is based off of it and Ken's Gated Comparator).

There's another very very, very spiffy 4015/3900 based random voltage function on the 208/Music Easel - I'm not sure if it's actually called it (it may be), but I refer to it as the "Uncorrelated Random Voltage" function - four outputs, each generating a different random voltage with each incoming clock pulse. Very cool, and very easy to build, once it's isolated from the 208 module.


Cheerios,
Scott

[loss1234]

eric

sorry i flaked out and forgot you had already posted links

scott-you mention a section of the easel that is easy once isolated...could you be a little more specific about which section?

thanks

[Scott Stites]

Loss,

It's on the 208 schematics page labled "Random Voltage Source Keyboard Interface".

By "isolate", I meant that it was linked into the programmer card circuitry through CD4016s. I separated it as a stand-alone project. Here's my drawing (which I built from). Note that the trigger pulse is squared off by a single section of a CD4016. I don't recall ever sucessfully triggering it from anything but that arrangment, though I think I probably could figure it out nowadays. That was fine for the time though - I used a section of an existing CD4016 that was being used in another circuit. I wouldn't try the circuit without doing it the same way, then devising another way to trigger the circuit once you know you have it going (unless you don't mind just using one section of a CD4016).

The original Buchla schematic has some relevant test notes on it.

Cheerios,
Scott

---

Buchla208_RandV.PNG
Description:	Buchla Random Voltage Function from 208/Music Easel (four random outputs from single 4015)
Filesize:	51.9 KB
Viewed:	1307 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[Scott Stites]

Probably not the right thread, let alone forum, but WTH. Here's my redraw of the absolute coolest S&H I've ever worked with - the one from the 266 SOU. Now, as with all Buchla CVs, it only handles positive voltages, but, my my - how it handles them!

It takes a single sample input, and outputs each sample out on the CV output (like a "normal" sample and hold). It has two additional CV outputs that alternate between samples - sample 1 goes out CV1, sample 2 goes out CV2, sample 3 goes out CV1, sample 4 goes out CV2, and so on. Plus, it provides alternating pulse outputs for these two CVs. I can't begin to describe the bug (and non-bug) music I've made just using this bad boy as a controller.

One of my favorite circuits of all time......

Cheerios,
Scott

---

Buchla266_S&H.PNG
Description:	Redrawn Buchla 266 SOU S&H Function
Filesize:	43.27 KB
Viewed:	1165 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[Scott Stites]

My redraws of the Quantized Random Voltage and Stored Random Voltage functions from the SOU. I noticed on my redraw of the SRV, the reset pins of the CD4015s were left floating (pins not shown on the original Buchla schematic). Pretty sure they should be grounded, so I grounded them.

I redrew the schematics in an attempt to better understand them and breadboard from them. The functions are so complex (at least to me) I'm not sure a redraw makes anything any more clear. They are masterworks of randomness with the tools at hand at the time, IMHO. They really make one think.

I've got the fluctating random voltage and voltage controlled integrator around somewhere, I believe. In the case of the fluctuating random voltage, I did modify the circuit to eliminate the hard-to-find MM device used for digital noise. I also added droopy S&H outputs to the FRV. Mebbe someday I'll dig them up.

Cheerios,
Scott

---

Buchla266_QRV.PNG
Description:	Redrawn Quantized Random Voltage from Buchla 266 SOU
Filesize:	72.24 KB
Viewed:	965 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

---

Buchla266_SRV.PNG
Description:	Redrawn Stored Random Voltage function from the Buchla 266 SOU
Filesize:	66.47 KB
Viewed:	971 Time(s)
This image has been reduced to fit the page. Click on it to enlarge.

---

[loss1234]

wow

these are just what i have been looking for!!

thanks so much

what a great fathers day gift!!

i will definitely be buildng these

[CJ Miller]

!!!!

I can entirely relate to being overwhelmed by Buchla's schematics, and ideas. I can hardly understand more pedestrian electronics, nevermind this...

I am on a big "stepped and smooth" kick right now, making my first Lunetta and also starting a portable synth. This is just the sort of stuff I was just dreaming about. I haven't even begun to try looking at the Easel schematics yet. Did you ever see the giant day-glo poster they did of the "Easel Weasel"?

I'm off to try getting some actual work/play done. Thank you so much for your redraws and ideas, Scott!

[loss1234]

scott

when you talk of the 4016 do you mean ICu6? this was part of the circuit originally? and i can just input a gate (high or low-square is what is meant by pulse right?)

and since the 4016 can take any kind of logic or signal in (basically) this is a pretty easy thing to interface with right?

this is amazing info

[loss1234]

oh and if you find

the fluctating random voltage and voltage controlled integrator redraws, well by all means i think we would ALL be interested!!

i cant believe you have been keeping this to yourself

[loss1234]

what do you normally use for the RANDOM PULSE IN?

would a noise gen put through a comparator work?

[Scott Stites]

"Random Pulse In" is just a label - the pulse itself need not be random - it's literally "pulse to generate the next random voltage".

Good question on the levels though - the Buchla idea of pulse levels and the Moog or Electronotes idea are quite divergent. You'll want to have pulses well in excess of 10V - I'm thinking at least 12V(?) to get some of these circuits to do their thang, unless you modify.

I've found my redraws of the Fluctuating Random Voltage function, but they're not in any shape for posting right now - they were a series of exchanges I had back and forth with Jeff when he was building his SOU adaptation. As I mentioned before, I modified a what now appears to me to be crude (but quite functional) replacement of the MM noise device used. I used transistor noise into a comparator to emulate the digital levels of the circuit. It looks goofy now, but the results were quite good. There are several add-ons I did at the time (S&H and trigger outputs for it) that I would be loathe to do now. I guess I've moved past "guilding the lily" . Still, the additions did work out quite well for me. I'll try to see if I can organize them before next week (I'll be out of country between June 23 and July 25).

Back to the RFV - what a cool, cool function. I think the Woggle is something like it. In the 266, it uses noise to frequency modulate a slow triangle wave (IIRC, the 265 used a sawtooth and transistor noise??). Anyway, the signal is applied as the input to a simple sample and hold circuit within the RFV function. The really cool thing is what Mr. Buchla did with the sample and hold output - he slewed it, but he devised the time constant of the slew to be directly tied to the time constant of the clock.

With a "normal" sample and hold lagged output, the time constant is fixed. So, if you set it for a high slew rate, the range of voltages it will cover are reduced by the frequency of the clock, more or less. In other words, if you were sampling the same signal with a high slew rate at a low sample rate, the slew would allow enough time for the output to rise to highest and lowest voltages. As the clock signal is increased, the voltage range becomes constricted - there isn't enough time to, say, go from a high sampled voltage to a subsequently low sampled voltage before the next sample comes through. At a low slew rate with a slow clock, the sample will "jump" from sample to sample - not abruptly, but not smoothly, either. It will linger at one level then "zoom" quickly to the next sampled level. The idea is that you want the slow rate (low probability of change) to be just as smooth in its rate of change as the fast rate (high probability of change), and vice-versa.

The FRV, however adjusts the slew rate along with the sample rate - these two parameters are what the "Probablity of Change" control effects. At low sample rates, the slew is set for a longer time constant - it will slowly glide from one sample to the next. At a high clock frequency, the slew rate is set for a shorter time constant, and proportionately glides quickly from one level to the next. The upshot is that the output is both smooth and non-restricted in the range it will fluctuate. It's a marvelous function.

Cheerios,
Scott

[ericcoleridge]

[ericcoleridge]

[yusynth]

Hi Scott

Thanks for sharing your redrawing of these Buchla modules, it saves us a lot of headaches indeed.
This is very much appreciated.

Cheers

Yves

[loss1234]

so if buchlas need a 12 volt pulse, does that mean a level shifted square wave that has its edge at 12volts? and if this is the case, how does CMOS handle such a high high?

(or can you just level shift the pulse POST cmos? as you can see i am quite confused and maybe i am totally misunderstanding what you meant. but from what you described, that would be one loud pulse~~!!

thanks

[funkyfarm]

[yusynth]

[funkyfarm]

oh yes, for sure.
thank you. (what a module !)

as my old Mac, i ran out of memory